Seasonal variability and trends in coastal upwelling across the Northwest African coastline, 1981-2012

The evolution of coastal upwelling systems under global warming is of huge biological, climatological and socio-economic importance. It was hypothesized in the early 1990's that upwelling might increase as a result of stronger land-ocean pressure gradients, however published results since then have conflicted as to whether coastal upwelling intensity has increased across the NW African coastline. Here, we present seasonal upwelling estimates for NW Africa (11-35°N) from 1981-2012, focusing mainly on changes during summer (JJA). Seasonal coastal upwelling indices are derived or estimated from several near-surface wind, sea-surface temperature (SST), sea-surface height (SSH), geometric vertical velocity and near-surface horizontal current datasets. For the wind-stress upwelling indices we use observational data (ICOADS), forecast model data (PFEL) and five atmospheric reanalysis datasets (ERA-Interim, NCEP-DOE II, 20th Century Reanalysis, MERRA and CFSR). For the SST indices we use the HadISST, Reynolds OISST and ICOADS datasets and for SSH we use satellite altimetry data from AVISO. Ocean reanalysis products (ORS4A, SODA 2.1.6 and GODAS) are used for ocean motion data. The numerous indices generally correlate well spatially, all reflecting a similar pattern induced by the trade wind climatology A statistically significant upwelling increase above 21°N is found in several indices, with a corresponding decrease in upwelling intensity below 20°N as well. This supports the upwelling intensification hypothesis as at approximately 20°N the summer trade winds are displaced by onshore monsoonal winds, which favour downwelling. However, these trend directions aren't ubiquitous across all the datasets, with most of the reanalysis wind indices suggesting no significant trend direction and sea-level height from altimetry showing no signs of lowering at coastal gridpoints (which would potentially indicate an upwelling increase). We attempt to validate if upwelling trends are real by using additional resources such as ocean colour data (SEAWIFS), scatterometer wind speeds (QuickSCAT), bias-adjusted wind measurements (WASWIND) and homogenised wind speeds from near-coastal weather stations. Fisheries landing data suggests a strong upwelling increase but isn't necessarily a pure climate signal. Internal modes of variability (NAO, EA, ENSO, AMO) are shown to weakly correlate with upwelling indices, but cannot explain much of the (summer) variation. We surmise that further warming is necessary for upwelling intensification to be fully realised, although the effects of coastal upwelling (i.e. negative coastal SST trends across certain areas) are already strongly recognized.